* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download molecular phylogeny of the haplosporidia based on
Survey
Document related concepts
Cre-Lox recombination wikipedia , lookup
Gene expression wikipedia , lookup
Transcriptional regulation wikipedia , lookup
Ancestral sequence reconstruction wikipedia , lookup
DNA barcoding wikipedia , lookup
Gene regulatory network wikipedia , lookup
Genome evolution wikipedia , lookup
Gene desert wikipedia , lookup
Non-coding DNA wikipedia , lookup
Gene expression profiling wikipedia , lookup
Promoter (genetics) wikipedia , lookup
Endogenous retrovirus wikipedia , lookup
Silencer (genetics) wikipedia , lookup
Point mutation wikipedia , lookup
Molecular evolution wikipedia , lookup
Transcript
J. Parasitol., 90(5), 2004, pp. 1111–1122 q American Society of Parasitologists 2004 MOLECULAR PHYLOGENY OF THE HAPLOSPORIDIA BASED ON TWO INDEPENDENT GENE SEQUENCES Kimberly S. Reece, Mark E. Siddall*, Nancy A. Stokes, and Eugene M. Burreson Department of Environmental and Aquatic Animal Health, Virginia Institute of Marine Science, The College of William and Mary, Gloucester Point, Virginia 23062. e-mail: [email protected] ABSTRACT: The phylogenetic position of the Haplosporidia has confounded taxonomists for more than a century because of the unique morphology of these parasites. We collected DNA sequence data for small subunit (SSU) ribosomal RNA and actin genes from haplosporidians and other protists for conducting molecular phylogenetic analyses to help elucidate relationships of taxa within the group, as well as placement of this group among Eukaryota. Analyses were conducted using DNA sequence data from more than 100 eukaryotic taxa with various combinations of data sets including nucleotide sequence data for each gene separately and combined, as well as SSU ribosomal DNA data combined with translated actin amino acids. In almost all analyses, the Haplosporidia was sister to the Cercozoa with moderate bootstrap and jackknife support. Analysis with actin amino acid sequences alone grouped haplosporidians with the foraminiferans and cercozoans. The haplosporidians Minchinia and Urosporidium were found to be monophyletic, whereas Haplosporidium was paraphyletic. ‘‘Microcell’’ parasites, Bonamia spp. and Mikrocytos roughleyi, were sister to Minchinia, the most derived genus, with Haplosporidium falling between the ‘‘microcells’’ and the more basal Urosporidium. Two recently discovered parasites, one from abalone in New Zealand and another from spot prawns in British Columbia, fell at the base of the Haplosporidia with very strong support, indicating a taxonomic affinity to this group. Haplosporidia is composed of histozoic and coelozoic parasites in a variety of freshwater and marine invertebrates; some species are significant pathogens of commercially important molluscs (Burreson et al., 2000). Since the discovery of the first species in the late 1800s, the Haplosporidia have been a troublesome group for taxonomists and phylogeneticists. Historically, the taxon has been treated as a last resort for a diversity of spore-forming parasites that have multinucleated naked cells (plasmodia) in their life cycles and that were not readily classifiable elsewhere (Sprague, 1979). There have been numerous taxonomic schemes proposed for placement of the group within the protists and for appropriate taxa within the Haplosporidia. Caullery and Mesnil (1899) established Haplosporidium for 2 parasites of marine annelids and placed the genus in the new order Haplosporidia in the class Sporozoa of the phylum Protozoa. Caullery (1953) recognized 6 genera in the order, and Kudo (1971) recognized 7 genera. A major change in the classification of the Haplosporidia was the separation of the Haplosporidia and the Paramyxea from other ‘‘sporozoa’’ by establishment of the new phylum Ascetospora (Sprague, 1979). This scheme proposed 2 classes: Stellatosporea, for the families Marteiliidae and Haplosporidiidae, and Paramyxea, for the family Paramyxidae. The family Haplosporidiidae contained only 3 genera, i.e., Haplosporidium, Minchinia, and Urosporidium. Desportes and Nashed (1983) recognized that the family Marteiliidae belonged in the class Paramyxea, not Stellatosporea, because of its development cycle, and proposed 2 classes in the phylum Ascetospora, i.e., Haplosporea and Paramyxea. However, they suggested that the 2 classes probably should be raised to phylum rank because of very different developmental sequences. Recently, the phylum Ascetospora has been abandoned, and Haplosporidia and Paramyxea have each been elevated to phylum rank (Desportes and Perkins, 1990; Perkins, 1990, 1991; Cavalier-Smith, 1993). Separate phylum rank for Haplosporidia and Paramyxea has been accepted by most researchers; however, Corliss (1994) re- Received 5 March 2003; revised 12 December 2003; accepted 17 February 2004. * Department of Invertebrate Zoology, American Museum of Natural History, New York, New York 10024. tained the phylum Ascetospora for both Haplosporidia and Paramyxea. The earliest molecular phylogenetic analyses for the Haplosporidia (Siddall et al., 1995; Flores et al., 1996) placed the phylum as a monophyletic group within the Alveolata and as a taxon of equal rank with the other alveolate phyla. A more recent analysis, with much more sequence data available for a variety of taxa, has placed the Haplosporidia as sister taxon to the Dictyosteliida (Berthe et al., 2000). The analysis by Berthe et al. (2000) also provided molecular phylogenetic support for separate phylum rank for Haplosporidia and Paramyxea. The most recent phylogenetic analysis involving the Haplosporidia (Cavalier-Smith and Chao, 2003) hypothesizes that the group includes the Paramyxea and falls within the Cercozoa. The phylum Haplosporidia was recently described by Perkins (2000) as a group of parasitic protists that form ovoid, walled spores with an orifice covered externally by a hinged lid or internally by a flap of wall material. There are 31 recognized species in 3 genera, i.e., Minchinia, Haplosporidium, and Urosporidium. However, recent molecular phylogenetic analyses supported the inclusion of the enigmatic genus Bonamia and Mikrocytos roughleyi, which are not known to form spores, within the Haplosporidia (Carnegie et al., 2000; CochennecLaureau et al., 2003). For this study, we used nucleotide sequence data of 2 genes, actin and the small subunit (SSU) ribosomal RNA (rRNA) gene, to investigate the position of the Haplosporidia among eukaryotes, to determine the taxonomic composition of the group, and to assess the monophyly of recognized genera. MATERIALS AND METHODS Sample collection Samples of haplosporidian-infected host tissues were collected in U.K., France, Spain, and Australia, and in Mississippi, Virginia, and Michigan in the United States (Table I). Cercozoan culture samples ATCC50317 and ATCC50318 were obtained from ATCC for DNA isolation and amplification of actin gene fragments. DNA isolation Spores of haplosporidian species were isolated after degradation of infected host tissue, and genomic DNA was extracted by methods previously described (Flores et al., 1996). DNA was isolated from the 1111 1112 THE JOURNAL OF PARASITOLOGY, VOL. 90, NO. 5, OCTOBER 2004 TABLE I. Collection sites and dates of samples collected for this study. Parasite species Haplosporidium nelsoni Host Sample collection site Minchinia teredinis Undescribed parasite Crassostrea virginica (Eastern oyster) Panopeus herbstii (mud crab) Crassostrea virginica (Eastern oyster) Helcion pellucidus (limpet) Physa parkeri (snail), Stagnicola emarginalus (snail) Microphallid trematode in Callinectes sapidus (blue crab) Stictodora lari (trematode) in Battilaria australis (whelk) Lepidochitona cinereus (chiton) Ruditapes decussatus (carpet shell clam) Teredo navalis (shipworm) Pandalus platyceros (spot prawn) Undescribed haplosporidian Cyrenoida floridana (marsh clam) Haplosporidium louisiana Haplosporidium costale Haplosporidium lusitanicum Haplosporidium pickfordi Urosporidium crescens Urosporidium sp. Minchinia chitonis Minchinia tapetis cercozoan cultures with the QIAamp DNA mini-kit (Qiagen, Carlsbad, California) according to the manufacturer’s protocol for tissue samples. Polymerase chain reaction amplification The SSU rRNA genes for most of the haplosporidian species were amplified in the polymerase chain reaction (PCR) using ‘‘universal’’ eukaryotic primers (Medlin et al., 1988) and conditions previously described (Flores et al., 1996). Internal SSU rRNA gene primers Urocomp280SSU (Reece and Stokes, 2003) and Uro280SSU (TAAYTGTTCGGATCGCATGGC) were paired with the eukaryotic primers 16S-A and 16S-B (Medlin et al., 1988), respectively, to amplify the Australian Urosporidium sp. rather than host DNA. Amplification conditions were same as for eukaryotic primers above. Generation of the SSU rRNA gene sequence for Haplosporidium pickfordi required several amplifications using specific primers: HAP-F1 (Renault et al., 2000) and 16S-B; 16S-A and MpickSSU391 (GCTTATTCAATCGGTAGGAGC); 16S-A and MpickSSU280 (CAATCGTCTATCCCCACTTG); Mpick125F (AACCGTGGTAACTCCAGGG) and Mpick1450R (TTATTGCCCCACGCTTCC); Mpick250F (CATTCAAGTTTCTGCCCTA TCAG) and Mpick1220R (TGTCTGGTAAGTTTTCCCGTGTTG); HAP-F1 and HAP-R3 (Renault et al., 2000). Reaction mixtures for PCR were same as above except with Mpick125F1 Mpick1450R and Mpick250F1 Mpick1220R, where buffer and polymerase were from the Expand High Fidelity PCR system (Roche Diagnostics, Indianapolis, Indiana). Cycling parameters consisted of an initial denaturation of 4 min at 94 C followed by 35 cycles of 30 sec at 94 C for denaturation, 30 sec at 48 C for HAP-F1116S-B and HAP-F11HAP-R3, 55 C for 16S-A1MpickSSU391 and 16S-A1MpickSSU280, 59 C for Mpick125F1Mpick1450R and Mpick250F1Mpick1220R for annealing, 1.5 min at 72 C for extension and a final extension of 5 min at 72 C. A central coding region of actin genes from the haplosporidians, their hosts, and cercozoans was amplified using various combinations of 2 forward (480, 481) and 2 reverse (482, 483) universal actin gene primers (Carlini et al., 2000) using amplification parameters and conditions previously described (Reece et al., 1997). Amplification products were cloned into either the plasmid vector pNoTA/T7 using the Prime PCR Cloner Cloning System (5 Prime-3 Prime, Inc., Boulder, Colorado) or into pCR2.1 using the TA Cloning Kit (Invitrogen, Carlsbad, California) according to the manufacturers’ instructions. DNA sequencing DNA clone inserts of SSU and actin gene fragments were sequenced manually as previously described (Reece et al., 1997) or by automated sequencing using either unidirectional or simultaneous bidirectional cy- Year collected Virginia 1992–1997 Gloucester Point, Virginia Wachapreague, Virginia 1991 1996 Cap de la Hague, France Douglas Lake, Michigan 1998 1999, 2000 Wachapreague, Virginia 1994 Sydney, New South Wales, Australia Plymouth, U.K. Galicia, Spain 2000 Wachapreague, Virginia Malaspina Strait, British Columbia, Canada Ocean Springs, Mississippi 1991, 1993, 1997 1999 1996 1996 1999 cle sequencing. Reactions for automated analysis were done as previously described (Reece and Stokes, 2003). Sequences were imported into MacVector 7.0 Sequence Analysis Software (Oxford Molecular Ltd., Oxford, U.K.) for trimming vector sequences, alignment, and for the actin gene sequences, coding region analyses. Actin gene introns (see Table II) were located, and the splice junction points were identified by using the MacVector 7.0 package to translate the DNA sequences in all 3 reading frames followed by coding region analyses and alignment to haplosporidian and other protistan actin gene fragments that did not contain introns. Sequence and phylogenetic analyses Occasionally, host or other contaminant SSU rRNA and actin genes were amplified rather than the targeted haplosporidian sequences. Identification of the DNA source for all amplified gene fragments involved BLAST (Altschul et al., 1990) searches of the National Center for Biotechnology Information (NCBI) GenBank database, as well as frequent intermediate phylogenetic analyses, which included previously obtained host, haplosporidian, and other protozoan sequences. In some cases, amplification of haplosporidian sequences from several infected host individuals was done to demonstrate that identical or highly similar (.98% sequence identity) sequences could be obtained from independent samples. Phylogenetic analyses were conducted in 2 series. The first concerned the assessment of the relationship of the Haplosporidia to other eukaryotic groups and comprised 121 taxa for the SSU rDNA sequences and 50 taxa for the actin sequences ranging across Eukaryota. Considerations at that level were limited principally to support for the groupings of and among various clades of protists, as opposed to support within each grouping. The second series concerned specific analysis of the phylogenetic relationships of genera and species within Haplosporidia using the closest relatives of the group, as determined from the first series of analyses. Sequences were aligned for each of these series independently. The SSU rRNA and actin gene sequences used in the phylogenetic analyses, along with their GenBank accession numbers, are listed in Table III. Sequences were aligned using the CLUSTALW algorithm (Thompson et al., 1994) in the MacVector 7.0 package using a variety of open and extend gap penalties in the ranges of 4–20 and 2–10, respectively. The SSU rRNA gene alignments were compared with secondary structure–based alignments done through the Ribosomal Database Project II website (http://rdp.cme.msu.edu/html/, Maidak et al., 2001). Final alignments used in the analyses were accomplished with gap penalties of 8 for insertions and 3 for extensions both in pairwise REECE ET AL.—MOLECULAR PHYLOGENY OF THE HAPLOSPORIDIA 1113 TABLE II. Actin gene introns, locations, and lengths. Sequences in bold were used in the phylogenetic analyses presented here. Parasite species No. of unique actin sequences No. of DNA identified clones sequenced Haplosporidium costale Haplosporidium louisiana 1 4 4 7 Haplosporidium nelsoni Minchinia chitonis 1 3 4 8 Minchinia tapetis 3 6 Minchinia teredinis 3 5 Urosporidium crescens Undescribed parasite CfP Undescribed parasite SPP 1 1 1 4 1 5 Gene designation No. of introns within amplified fragment HcAc1 HlAc1 HlAc5 HlAc8 HlAc11 HnAc3 McAc5 McAc11 McAc42 MtaAc1 MtaAc7 MtaAc9 MteAc2 MteAc5 MteAc9 UcAc1 CfhapAc64 SppAc6 2 0 0 0 3 0 1 3 1 0 2 0 2 2 0 0 1 4 Intron positions* (intron lengths in bp) 131, 172 (18, 17) 131, 172, 224 (177, 63, 83) 304 (124) 131, 172, 224 (25, 25, 28) 304 (114) 131, 172 (33, 77) 172, 224 (22, 60) 131, 172 (23, 25) 172 (25) 131, 172, 224, 304 (50, 60, 222, 185) * Intron positions are designated relative to the homologous amino acid position in the human aortic actin gene (GenBank NMp005159). and in multiple alignment phases. In addition, aligned sequences were analyzed in various combinations to determine the sensitivity of various groups to the relative inclusion and exclusion of potentially phylogenetically informative sites. In the first series of analyses, which concerned the relative position of Haplosporidia within Eukaryota, it was noted that there were large inserted regions in the SSU rDNA sequences of some taxa, including some of the haplosporidians, e.g., the prawn parasite and Urosporidium species. In light of this, a total of 798 sites in the alignment were targeted for removal to determine the effect that removal or inclusion of these sequences had on results. Similarly, in terms of the actin nucleotide data, the relative contribution of third codon positions to the predicted phylogenies was assessed, both through their exclusion and by using translated amino acid sequences. All combinations of SSU rDNA nucleotides, with and without the 798 sites targeted for exclusion, alone, and in combination with actin nucleotides, with and without third codon positions, or using translated amino acid sequences alone or in combination with the SSU rDNA data, were examined in terms of their relative support for higher-level relationships among the included eukaryotic taxa. Insofar as the specifics of species-level relationships were of only passing interest in the analyses of all Eukaryotic taxa, groupings and relative levels of support were determined through unweighted parsimony jackknife methods (Farris et al., 1996) for nucleotide combinations using XAC (Farris, 1998). Unweighted parsimony jackknife and bootstrap in PAUP 4.0b10 (Swofford, 2002) were used for those analyses concerning amino acid sequences with the ‘‘emulate Jac’’ option. In each case, 1,000 replicates were used, each with 5 random additions and with a subtree-pruningregrafting branch swapping algorithm. After determinations of jackknife support in these higher-level Eukaryota analyses, we also searched for optimal trees on the combined SSU rDNA and actin amino acid sequences because jackknife values from this data combination were intermediate among the 6 combinations tried. This search used 30 random taxon addition search sequences followed by branch breaking (TBR branch swapping). With respect to the secondary analyses in which we were concerned with the specifics of the relationships within Haplosporidia, sequences were realigned for the 26 included taxa for the combined analysis according to the same parameters as above. In these analyses, all sites for both nucleotide sets were included and combinations of data considered SSU rDNA alone, as well as those data in combination with actin nucleotides or with translated actin amino acid sequences. Thorough heuristic searches for most parsimonious trees (as opposed to jackknife support trees) were conducted with PAUP 4.0b10 (Swofford, 2002) for all the secondary data sets. In each case, searches involved 30 random taxon addition search sequences followed by branch breaking (TBR branch swapping). RESULTS From some samples, novel host as well as parasite SSU rRNA or actin gene sequences (or both) were identified. Host and parasite SSU rRNA and actin gene sequences that were confidently identified as either of host or parasite origin by BLAST searches and phylogenetic analyses in this study were deposited in GenBank. Additional sequences whose origin could not be determined were obtained from many samples and were presumed to be from contaminating organisms. Jackknife values obtained from the various data combinations used across Eukaryota for determination of the closest extant relatives of Haplosporidia are listed in Table IV. Bootstrap values were comparable with the jackknife values in all analyses, never varying by more than 3%, and usually bootstrap support values were slightly higher. With all but 1 of the data set combinations, the Haplosporidia were supported as having arisen from a common ancestor with the Cercozoa. With the complete combined nucleotide data set, which included all the SSU rRNA and actin gene nucleotides, the Haplosporidia were not supported as having arisen from a common ancestor with the Cercozoa. This complete nucleotide data set, however, failed to find greater than 50% support for the Haplosporidia grouping with any of the other Eukaryotic clades used in this analysis. Parsimony analysis with the actin amino acid data set alone grouped the Haplosporidia with the Foraminifera and most closely to the type 1 foraminiferan actins (Keeling, 2001). The foraminiferan and haplosporidian clade was sister to the cercozoans. The support for these groupings, however, was below 50%. All the remaining data sets found that Haplosporidia and Cercozoa are clades with support levels ranging from 61% with the SSU rDNA alone up to 81% with the SSU rDNA sequences 1114 THE JOURNAL OF PARASITOLOGY, VOL. 90, NO. 5, OCTOBER 2004 TABLE III. Taxa included in phylogenetic analysis and the GenBank accession numbers of their SSU rRNA and actin gene sequences. Sequences determined for this study are shown in bold. Taxon SSU rRNA gene Actin gene Bonamia ostreae Bonamia sp. Haplosporidium costale Haplosporidium louisiana Haplosporidium lusitanicum Haplosporidium nelsoni Haplosporidium pickfordi Mikrocytos roughleyi Minchinia chitonis Minchinia tapetis Minchinia teredinis Urosporidium crescens Urosporidium sp. Undescribed Cyrenoida floridana parasite (CfP) Undescribed Haliotis iris parasite (NZAP) Undescribed Pandalus platyceros parasite (SPP) Allogromia sp. Ammonia sp. Ammonia beccarii Reticulomyxa filosa Cercomonas longicauda Cercomonas sp.—ATCC50317 Cercomonas sp.—ATCC50318 Chlorarachnion reptans Chlorarachnion sp. Euglypha rotunda Heteromita globosa Massisteria marina Unidentified cercozoan species Phagomyxa bellerocheae Phagomyxa odontellae Plasmodiophora brassicae Achlya bisexualis Bolidomonas pacifica Cafeteria roenbergensis Cafeteria sp. Chrysosaccus sp. Chrysosphaera parvula Chrysoxys sp. Ciliophrys infusionum Costaria costata Fucus vesiculosus Hyphochytrium catenoides Lagynion scherffelii Ochromonas danica Phytophthora megasperma Phytophthora infestans Rhizidiomyces apophysatus Skeletonema costatum C9G Labyrinthula sp. QPX Ulkenia profunda Ajellomyces capsulatus Apusomonas proboscidea Candida glabrata Kluyveromyces lactis Neurospora crassa Pneumocystis carinii Saccharomyces cerevisiae Acanthamoeba castellanii AF192759, AF262995 AF337563 AF387122 U47851 AY449713 U19538 AY452724 AF508801 AY449711 AY449710 U20319 U47852 AY449714 AY449712 AF492442 AY449715–716 X86093† NA U07937† AJ132367† AF101052 U42449 U42450 U03477 NA X77692 U42447 AF174374 UEU130858 AF310903† AF310904† U18981† M32705 AF167154 AF174364 AF174366 AF123300 AF123299 AF123302 L37205 X53229 U97110 AF163294 AF123288 M32704 X54265 NA AF163295 X85395 AF474172 AB022105 AF261664 L34054 Z75307 L37037 X51831 X51830 X04971 S83267 J01353 AF251938 NA* NA AY450407 AY450408–411 NA AY450412 NA NA AY450413–415 AY450416–418 AY450419–421 AY450422 NA AY450423 NA AY450424 AJ132370–371 AJ132372–373 NA AJ132374–375 NA AF363534 AF363536 NA AF363528 NA NA NA NA NA NA NA X59936 NA NA NA NA NA NA NA X59937 X98885 NA NA NA NA M59715 NA NA NA NA NA NA U17498 NA AF069746 M25826 U78026 L21183 V01288 V00002, J01016 REECE ET AL.—MOLECULAR PHYLOGENY OF THE HAPLOSPORIDIA TABLE III. Continued. Taxon SSU rRNA gene Actin gene Chilomonas paramecium Chlamydomonas reinhardtii Emiliania huxleyi Nitella flexilis Chondrus crispus Stylonema alsidii Marteilia refringens Alexandrium fundyense Amphidinium carterae Gonyaulax spinifera Hematodinium sp. Lepidodinium viride Pfiesteria piscicida Prorocentrum minimum Symbiodinium corculorum Symbiotic dinoflagellate BBSR 323 Perkinsus chesapeaki Perkinsus marinus Perkinsus olseni Babesia bovis Caryospora bigenetica Colpodella sp. Cryptosporidium parvum Eimeria tanella Parvilucifera infectans Sarcocystis hominis Theileria parva Toxoplasma gondii Anophyroides haemophila Didinium nasutum Entodinium caudatum Euplotes crassus Frontonia vernalis Oxytricha nova Oxytricha trifallax Paramecium tetraurelia Prorodon viridis Strombidium purpureum Tetrahymena pyriformis Tetrahymena thermophila Acanthometra sp. Arthracanthid 206 Haliommatidium sp. Encephalitozoon cuniculi Encephalitozoon hellum Entamoeba histolytica Euglena gracilis Trichomonas foetus Trichomonas vaginalis Trypanosoma brucei Haliotis iris Crassostrea virginica Cyrenoida floridana Helcion pellucidus Lepidochitona cinereus Pandalus platyceros Physa parkeri Ruditapes decussatus Stictodora lari Teredo navalis L28811 M32703 M87327 U05261 Z14140 AF168633 AJ250699† U09048 AF009217 AF052190 AF286023 AF022199 AF077055 Y16238 L13717 U52356 AF042707 AF042708 L07375 L19078 AF060975 AY142075, AY449717 AF115377 U40264 AF133909 AF006470 L02366 U03070 U51554 U57771 U57765 AY007438 U97110 X03948 AF164121 X03772 U97111 U97112 M98021 X56165 AF063240 AF063239 AF018159 L17072 M87327 X65163 M12677 M81842 U17510 AJ009141 AF492441 L78851† NA NA NA NA NA AF29512† NA NA NA D50839 S64188 NA U03676 NA NA NA U84289 NA NA NA NA U84290 NA NA NA U84287 NA NA NA NA M86241 NA NA NA NA U10429 NA NA AF078106 J04533 NA M22480 NA AF043608 NA NA X05195 M13939 NA NA NA NA AF031701 M19871 AF057161 NA AF237734 M20310 NA X75894† AY452511–512† AY452513† AY452514–515† AY452516† AY485722† AY452517–518† AY452519–521† AY452522† * NA, sequence not available. † Sequences used for analyses, but not included in trees presented. 1115 1116 THE JOURNAL OF PARASITOLOGY, VOL. 90, NO. 5, OCTOBER 2004 TABLE IV. Summary of jackknife support values for phylogenetic analyses with various data sets.* Ciliophora Dinozoa Apicomplexa Alveolata Stramenopiles Labyrinthulids Fungi Chlorophytes (FUN 1 CHL) Haplosporidia Cercozoa (HAP 1 CERC) SSU SSUX SSU1ACT SSUX1ACTX SSU1ACTX SSUX1ACT SSU1ACTaa SSUX1ACTaa 100 100 — 52 67 100 97 90 78 91 100 61 99 95 — 73 69 91 88 86 77 67 100 64 95 100 — 54 80 96 93 97 — 83 100 —† 98 94 — 78 77 86 72 — 53 90 100 79 100 100 — 64 78 94 92 74 68 94 100 63 88 96 — — 75 97 68 91 — 92 99 75 100 100 — 70 72 94 98 86 79 95 100 74 99 97 — 77 75 94 84 71 66 87 100 81 * SSU, all SSU rDNA nucleotides aligned by CLUSTALW; SSUX, SSU rDNA nucleotides aligned by CLUSTALW, with 798 poorly aligned nucleotide positions removed; ACT, all actin gene nucleotides aligned by CLUSTALW; ACTX, actin gene nucleotides aligned by CLUSTALW, with the third positions of codons removed. † No alternative grouping of the Haplosporidia was supported by this analysis. combined with actin amino acid sequences but excluding the 798 sites corresponding to SSU rDNA insertions in the haplosporidian sequences. Intermediate between these values were the analyses using all SSU rDNA data and either the actin amino acid sequences with 74% support for the cercozoan–haplosporidian clade or the actin nucleotide sequences with third positions excluded with 63% support for the cercozoan–haplosporidian clade. Overall, both these analyses rendered high levels of support for other known eukaryotic groups (Table IV). The SSU rDNA data combined with the actin amino acid sequences, when subjected to a full parsimony search, yielded 4 trees with equal length of 17,977, a consistency index (CI) of 0.306, and a retention index (RI) of 0.557, the strict consensus of which is illustrated in Figure 1. Foraminifera SSU rRNA gene sequences could not be confidently aligned with the other sequences and, therefore, were not used in these analyses. The consensus tree of 122 equally parsimonious trees with a length of 1,153 (CI 5 0.440; RI 5 0.555) resulting from analysis with the actin amino acid data set is shown in Figure 2. For parsimony analyses examining the in-group relationships of the Haplosporidia, 8 cercozoan taxa were used as the outgroup (Fig. 3). Use of SSU rDNA data alone yielded a single tree with length 3,838 (CI 5 0.600; RI 5 0.658). Use of SSU rDNA in combination with translated actin amino acids also rendered a single tree, identical in topology to the SSU rDNA tree, with length 4,172 (CI 5 0.607; RI 5 0.649). Use of SSU rDNA in combination with actin nucleotides yielded 2 trees each with length 5,213 (CI 5 0.572; RI 5 0.589). The consensus of those 2 trees disagreed with the other analyses in moving a clade composed of H. pickfordi and Haplosporidium lusitanicum to a sister group position with Haplosporidium louisiana as opposed to with Haplosporidium costale. The undescribed haplosporidian parasite from Cyrenoida floridana consistently grouped within the described Minchinia species. All combined data set analyses agreed on a variety of other findings including that (1) Minchinia is monophyletic (if the parasite from C. floridana is considered a Minchinia species); (2) Urosporidium is monophyletic; (3) the nonspore-forming ‘‘microcell’’ parasites within the Haplosporidia, e.g., Bonamia spp. and M. roughleyi, are monophyletic and sister to Minchinia; (4) Haplosporidium as currently constituted is paraphyletic; and (5) the earliest lineages in Haplosporidia are the yet to be described parasite species recently obtained from prawns and from abalone. In several cases, more than 1 actin gene type was amplified from DNA of a haplosporidian species. Multiple paralogs were found in H. louisiana, Minchinia chitonis, Minchinia teredinis, and Minchinia tapetis. Only a single type of actin gene was identified from Haplosporidium nelsoni, H. costale, and Urosporidium crescens, and no actin gene that could be confidently identified as a parasite sequence was successfully amplified from H. lusitanicum, H. pickfordi, or from the Australian Urosporidium sp. found in Stictodora lari (trematode) of Battilaria australis (whelk), although 10–27 DNA clones were sequenced from each of these samples. As sequences were obtained, phylogenetic analyses were conducted to identify orthologous genes for inclusion in the overall phylogenetic analyses. Many haplosporidian actin gene sequence fragments were found to contain introns. Introns were found in at least 1 actin gene paralog from each of the haplosporidian taxa from which actin genes were able to be amplified, except from H. nelsoni, and U. crescens. Orthologs and paralogs of parasite origin were identified by phylogenetic analyses with known actin gene sequences from hosts and other protozoans. The genes with introns as well as the single intronless actin genes isolated from H. nelsoni and U. crescens were determined to be orthologous (Table II) and were used in the comprehensive phylogenetic analyses. Intron positions were conserved among the haplosporidian actin genes, with 4 different intron positions identified at positions corresponding to the amino acids 131, 172, 224, and 304 of the human aortic actin gene (GenBank NMp005159). The total number of introns within the amplified gene fragment ranged from 1 to 4 (Table II). Individual actin genes from the described haplosporidian species contained 1–3 introns, and the introns were found at all 4 of the positions in the actin gene of the haplosporidianlike spot prawn parasite. Intron lengths ranged from very short introns of 17–33 nucleotides, up to 222 nucleotides. The introns were flanked by typical splice junction sequences with ‘‘GT’’ at the 59 end and ‘‘AG’’ at the 39 end of all introns. REECE ET AL.—MOLECULAR PHYLOGENY OF THE HAPLOSPORIDIA 1117 FIGURE 1. Strict jackknife consensus of 4 equal length trees resulting from parsimony analysis with the SSU rDNA and actin amino acid data set. Analysis was done on the complete taxonomic data set with 798 poorly aligned nucleotide positions in the SSU rDNA removed. Jackknife support values are given at the nodes. Dashed lines indicate clades that did not have jackknife support values above 50. 1118 THE JOURNAL OF PARASITOLOGY, VOL. 90, NO. 5, OCTOBER 2004 FIGURE 2. Strict jackknife consensus of 122 equal length trees resulting from parsimony analysis with the actin amino acid data set. Analysis was done on actin sequences from 47 taxa. Jackknife support values are given at the nodes. Dashed lines indicate clades that did not have jackknife support values above 50. REECE ET AL.—MOLECULAR PHYLOGENY OF THE HAPLOSPORIDIA 1119 FIGURE 3. Strict jackknife consensus trees of parsimony analyses conducted with the reduced taxonomic data set that included the cercozoans and haplosporidians for examination of relationships within the phylum Haplosporidia. Jackknife support values are given at the nodes. A. Tree resulting from analysis of the SSU rDNA sequence data. Tree statistics: length (L) 5 3,838, consistency index (CI) 5 0.600, retention index (RI) 5 0.658. B. Strict consensus of 2 trees of equal length resulting from analysis of the SSU rDNA and actin gene sequence data. Tree statistics: L 5 5,213, CI 5 0.572, RI 5 0.589. C. Tree resulting from analysis of the SSU rDNA and actin amino acid sequence data. Tree statistics: L 5 4,172, CI 5 0.607, RI 5 0.649. 1120 THE JOURNAL OF PARASITOLOGY, VOL. 90, NO. 5, OCTOBER 2004 DISCUSSION This combined actin–SSU rRNA data set, including the data for several haplosporidian taxa collected during this study and additional data already in GenBank, places Haplosporidia in a position of recent common ancestry with cercozoans and not within the Alveolata as previously hypothesized on the basis of the SSU rRNA gene data alone (Siddall et al., 1995; Flores et al., 1996). This study includes not only an additional gene and many additional haplosporidian taxa but also several more protistan SSU rRNA and actin gene sequences that were not available for those studies. The haplosporidians grouped with the foraminiferans and cercozoans in analyses based on actin sequences alone, although jackknife support for this relationship was below 50% (Fig. 2). We did not, however, include the foraminiferan SSU rDNA sequences in any of the analyses because we did not have confidence in alignments that included these divergent sequences. Keeling (2001) found that the Foraminifera and Cercozoa were closely related in actin and tubulin gene phylogenies. Any more definitive conclusions regarding the relationships between Haplosporidia and Foraminifera, however, will require additional data. Although previous studies have suggested that Haplosporidia and Cercozoa are closely related (Cavalier-Smith, 2002; Cavalier-Smith and Chao, 2003), this is the first study to demonstrate moderate support (JK 5 74) for this affinity. Cavalier-Smith (2002) stated that the Haplosporidia are members of the Cercozoa but presented absolutely no data on which to evaluate this claim. Cavalier-Smith and Chao (2003) had only weak support (BS 5 20 with Marteilia refringens included in the analysis and BS 5 60 with it removed) for inclusion of the Haplosporidia within the Cercozoa. Although that study also suggested that M. refringens was related to the haplosporidians, the results of this study were consistent with the results of Berthe et al. (2000), in that we found no support for this relationship in any of our analyses. The fact that there is strong support in this study for many other accepted taxonomic groupings, including the stramenopiles, alveolates, labyrinthulids, fungi, chlorophytes and cercozoans (see Table IV), lends support to the validity of these results. Moreover, the use of combined data from these 2 loci provided more support for grouping Haplosporidia with Cercozoa, than either locus did alone and with support values similar to those seen for Alveolata and Stramenopiles arising at comparable depths in the resulting tree. In addition, if the Cercozoa are accepted as a phylum, then the results of this study hypothesize that the Haplosporidia are also a distinct phylum, rather than included within the Cercozoa as hypothesized by Cavalier-Smith and Chao (2003). This phylogenetic study based on molecular data from 2 independent DNA sequences, namely SSU rRNA and actin gene sequences, generally supports the earlier taxonomic assignments within the group Haplosporidia that were largely based on spore morphologies (Sprague, 1979; Ormières, 1980; McGovern and Burreson, 1990; Perkins, 1990, 1991, 2000). Monophyly of Urosporidium and Minchinia was supported in this study. Haplosporidium is paraphyletic and falls intermediate to the more basal Urosporidium and the most derived genus Minchinia, as previously hypothesized by Ormières (1980). Many of the more recent observations regarding taxa found to have haplosporidian affinities and reassignments within the group were supported by this study. The Bonamia spp. and M. roughleyi fell within the Haplosporidia, as had been observed in previous studies (Carnegie et al., 2000; Cochennec-Laureau et al., 2003). These ‘‘microcells’’ grouped together in this analysis as a sister group to Minchinia. The results of this study strongly suggest that the undescribed C. floridana parasite is a Minchinia species because it fell between M. chitonis and M. teredinis within a monophyletic Minchinia clade in all the analyses. In another study, the freshwater parasite Minchinia pickfordae was assigned recently to Haplosporidium on the basis of spore ornamentation (Burreson, 2001) and renamed H. pickfordi. Its removal from Minchinia was supported here. A haplosporidianlike abalone parasite (Diggles et al., 2002; Hine et al., 2002; Reece and Stokes, 2003) fell at the base of the Haplosporidia, sister to another haplosporidianlike parasite recently identified in the spot prawn, Pandalus platyceros (Bower and Meyer, 2002). Overall, there was strong support (jackknife support value 5 95) for monophyly of the group Haplosporidia, provided that Bonamia spp. and M. roughleyi are included. There was also strong support (jackknife support value 5 95) for the placement of the abalone and spot prawn parasites at the base of the haplosporidian clade. It is interesting to note that introns were found within actin genes from many of the haplosporidian species in this study. The presence and the position of introns often lent further support to results of the phylogenetic analyses. Four intron positions were identified within the haplosporidian actin genes, with the number of introns within a particular gene fragment ranging from 1 to 3 within genes from the described species. At least 1 actin gene paralog with introns was isolated from each of the Minchinia species. The actin gene from the undescribed C. floridana parasite contained an intron, consistent with its placement within this genus. Introns were also found in the spot prawn parasite actin gene at each of the 4 conserved positions found in haplosporidian actin genes, lending additional support to the phylogenetic affinity of this spot prawn parasite with Haplosporidia. Several of these introns were surprisingly short (17–33 bp). Even these short introns, however, had characteristic splice junction sequences and began at their 59 ends with the dinucleotide GT and ended at their 39 ends with the dinucleotide AG. In addition, when the intron sequences were removed, the remaining exon sequences could be translated in frame to yield typical actin amino acid sequences, further supporting the validity of even the shortest introns. Introns were also found in actin genes from a cercozoan species (Keeling, 2001), the group found to be sister to Haplosporidia in the combined data set analysis, although the intron positions differed from those seen in the haplosporidian actin genes. To our knowledge, actin genes have not yet been isolated from Bonamia spp., M. roughleyi, or the abalone parasite, and although we attempted, we did not successfully isolate actin genes from the Australian Urosporidium sp., H. pickfordi, or H. lusitanicum. It is quite likely that several actin gene paralogs from these parasites have yet to be isolated. Phylogenetic analyses with these sequences as well as characterization of the intron structure may lend additional insight into relationships within the group. Overall, the results of this molecular study raise interesting questions regarding spore formation for taxa within the group Haplosporidia. Spores have not been observed for either the spot prawn or abalone parasites that lie at the base of the clade REECE ET AL.—MOLECULAR PHYLOGENY OF THE HAPLOSPORIDIA (Bower and Meyer, 2002; Diggles et al., 2002; Hine et al., 2002), suggesting that the ancestral state included a lack of the ability to form spores and that spore formation arose along the lineage to the Urosporidium, Haplosporidium, and Minchinia genera. Spores have not been observed, however, for the Bonamia spp. and M. roughleyi, which in the phylogenetic analyses fell between the spore-forming genera Minchinia and Haplosporidium. This hypothesizes a loss of the ability to form spores in the microcell group. It is possible, however, that the life stages of these undescribed basal parasites identified to date and the microcells are intermediate stages of previously undiscovered haplosporidians and that the spore stages have not yet been observed. Alternatively, the ancestor to the Haplosporidia had the ability to form spores, but this was lost in the lineages to the spot prawn and abalone parasites, as well as in the lineage to the microcells, whereas the other taxa have retained this ability. Relationships within the group Haplosporidia hypothesized by the molecular phylogenetic analyses provide a framework for assessing proposed relationships within the group based on morphology. Perkins (2000) assigned species to Minchinia if they had spore extensions that were visible with the light microscope and to Haplosporidium if the spore ornamentation was not visible with the light microscope. Most other researchers have followed Ormières’ (1980) criteria and assigned species to Minchinia if the spore ornamentation was composed of epispore cytoplasm and to Haplosporidium if the spore ornamentation was derived from the spore wall (McGovern and Burreson, 1990; Hine and Thorne, 1998; Azevedo et al., 1999; Burreson, 2001). The molecular phylogenetic analyses presented here support the importance of ornamentation origin. All the species included in the analysis that have ornamentation composed of epispore cytoplasm (Minchinia spp.) formed a monophyletic group. The species that have ornamentation derived from the spore wall (Haplosporidium spp.) formed a paraphyletic clade, suggesting that more genera are necessary to encompass the morphological diversity in species with spore wall–derived ornamentation. Unfortunately, at present, there are insufficient data on spore wall ornamentation for many of the haplosporidian species to propose new generic assignments. As knowledge increases, it will be interesting to assess the concordance between molecular and morphological data sets. ACKNOWLEDGMENTS The authors thank Brenda S. Kraus, Kathleen Apakupakul, and Karen Hudson for technical support and Antonio Villalba, Barbara Nichols, Susan Bower, Gary Meyer, Ben Diggles, and John Walker for providing samples. This work was supported by grant BIO-DEB-9629487 from the U.S. National Science Foundation. VIMS contribution 2617. LITERATURE CITED ALTSCHUL, S. F., W. GISH, W. MILLER, E. W. MYERS, AND D. J. LIPMAN. 1990. Basic local alignment search tool. Journal of Molecular Biology 215: 403–410. AZEVEDO, C., J. MONTES, AND L. CORRAL. 1999. A revised description of Haplosporidium amoricanum, parasite of Ostrea edulis L. from Galicia, northwestern Spain, with special reference to the sporewall filaments. Parasitology Research 85: 977–983. BERTHE, F. C. J., F. LE ROUX, E. PEYRETAILLADE, P. PEYRET, D. RODRIGUEZ, M. GOUY, AND C. P. VIVARÉS. 2000. Phylogenetic analysis of the small subunit ribosomal RNA of Marteilia refringens validates 1121 the existence of phylum Paramyxea (Desportes and Perkins, 1990). Journal of Eukaryotic Microbiology 47: 288–293. BOWER, S. M., AND G. R. MEYER. 2002. Morphology and ultrastructure of a protistan pathogen in the haemolymph of shrimp (Pandalus spp.) in the northeastern Pacific Ocean. Canadian Journal of Zoology 80: 1055–1068. BURRESON, E. M. 2001. Spore ornamentation of Haplosporidium pickfordi Barrow, 1961 (Haplosporidia), a parasite of freshwater snails in Michigan, USA. Journal of Eukaryotic Microbiology 48: 622– 626. ———, N. A. STOKES, AND C. S. FRIEDMAN. 2000. Increased virulence in an introduced pathogen: Haplosporidium nelsoni in the eastern oyster Crassostrea virginica. Journal of Aquatic Animal Health 12: 1–8. CARLINI, D. B., K. S. REECE, AND J. E. GRAVES. 2000. Actin gene family evolution and the phylogeny of coleoid cephalopods (Mollusca: Cephalopoda). Molecular Biology and Evolution 17: 1353–1370. CARNEGIE, R. B., B. J. BARBER, S. C. CULLOTY, A. J. FIGUERAS, AND D. L. DISTEL. 2000. Development of a PCR assay for detection of the oyster pathogen Bonamia ostreae and support for its inclusion in the Haplosporidia. Diseases of Aquatic Organisms 42: 199–206. CAULLERY, M. 1953. Appendice aux Sporozoaires. Classe des haplosporidies. In Protozoaires: Rhizopodes, Actinopodes, Sporozoaires, Cnidospories, P. P. Grassé (ed.). Traité de Zoologie, vol. I, no. 2. Masson et Cie, Paris, France, p. 922–934. ———, AND F. MESNIL. 1899. Sur le genera Aplosporidium (nov.) et l’ordre nouveau des Aplosporidies. Compte rendu des seances de la Societe de Biologie et de ses Filiales 51: 789–791. CAVALIER-SMITH, T. 1993. Kingdom Protozoa and its 18 phyla. Microbiological Reviews 57: 953–994. ———. 2002. The phagotrophic origin of eukaryotes and phylogenetic classification of Protozoa. International Journal of Systematic and Evolutionary Biology 52: 297–354. ———, AND E. E. CHAO. 2003. Phylogeny of Choanozoa, Apusozoa and other Protozoa and early eukaryote megaevolution. Journal of Molecular Evolution 56: 540–563. COCHENNEC-LAUREAU, N., K. S. REECE, F. C. J. BERTHE, AND P. M. HINE. 2003. Revisiting Mykrocytos roughleyi taxonomic affiliation leads to the genus Bonamia (Haplosporidia). Diseases of Aquatic Organisms 54: 209–217. CORLISS, J. O. 1994. An interim utilitarian (‘‘user-friendly’’) hierarchical classification and characterization of the protists. Acta Protozoologica 33: 1–51. DESPORTES, I., AND N. N. NASHED. 1983. Ultrastructure of sporulation in Minchinia dentali (Arvy), an haplosporean parasite of Dentalium entale (Scaphopoda, Mollusca): Taxonomic implications. Protistologica 19: 435–460. ———, AND F. O. PERKINS. 1990. Phylum Paramyxea. In Handbook of Protoctista, L. Margulis, J. O. Corliss, M. Melkonian, and D. J. Chapman (eds.). Jones and Bartlett Publishing, Boston, Massachusetts, p. 30–35. DIGGLES, B. K., J. NICHOL, P. M. HINE, S. WAKEFIELD, N. COCHENNECLAUREAU, R. D. ROBERTS, AND C. S. FRIEDMAN. 2002. Pathology of cultured paua (Haliotis iris) infected by a novel haplosporidian parasite, with some observations on the course of disease. Diseases of Aquatic Organisms 50: 219–231. FARRIS, J. S., V. A. ALBERT, M. KÄLLERSJÖ, D. LIPSCOMB, AND A. G. KLUGE. 1996. Parsimony jackknifing outperforms neighbor-joining. Cladistics 12: 99–124. ———. 1998. Xac. Program and documentation. Swedish Museum of Natural History, Stockholm, Sweden. FLORES, B. S., M. E. SIDDALL, AND E. M. BURRESON. 1996. Phylogeny of the Haplosporidia (Eukaryota: Alveolata) based on small subunit ribosomal RNA gene sequence. Journal of Parasitology 82: 616– 623. HINE, M., AND T. THORNE. 1998. Haplosporidium sp. (Haplosporidia) in hatchery-reared pearl oysters, Pinctada maxima (Jameson, 1901), in north Western Australia. Journal of Invertebrate Pathology 71: 48–52. HINE, P. M., S. WAKEFIELD, B. K. DIGGLES, V. L. WEBB, AND E. W. MAAS. 2002. Ultrastructure of a haplosporidian containing Rickettsiae, associated with mortalities among cultured paua Haliotis iris. Diseases of Aquatic Organisms 49: 207–19. 1122 THE JOURNAL OF PARASITOLOGY, VOL. 90, NO. 5, OCTOBER 2004 KEELING, P. J. 2001. Foraminifera and Cercozoa are related in actin phylogeny: Two orphans find a home? Molecular Biology and Evolution 18: 1551–1557. KUDO, R. R. 1971. Protozoology, 5th ed. Charles C. Thomas Publishing, Springfield, Massachusetts, 1,174 p. MAIDAK, B. L., J. R. COLE, T. G. LILBURN, C. T. PARKER JR., P. R. SAXMAN, R. J. FARRIS, G. M. GARRITY, G. J. OLSEN, T. M. SCHMIDT, AND J. M. TIEDJE. 2001. The RDP-II (Ribosomal Database Project). Nucleic Acids Research 29: 173–174. MCGOVERN, E. R., AND E. M. BURRESON. 1990. Ultrastructure of Minchinia sp. from shipworms (Teredo spp.) in the Western North Atlantic, with a discussion of taxonomy of the Haplosporidiidae. Journal of Protozoology 37: 212–218. MEDLIN, L., H. J. ELWOOD, S. STICKEL, AND M. L. SOGIN. 1988. The characterization of enzymatically amplified eukaryotic 16S-like rRNA-coding proteins. Gene 71: 491–499. ORMIÈRES, R. 1980. Haplosporidium parisi n. sp., Haplosporidie parasite de Serpula vermicularis L. étude ultrastructurale de la spore. Protistologica 16: 467–474. PERKINS, F. O. 1990. Phylum Haplosporidia. In: Handbook of Protoctista, L. Margulis, J. O. Corliss, M. Melkonian, and D. J. Chapman (eds.). Jones and Bartlett Publishing, Boston, Massachusetts, p. 19– 29. ———. 1991. ‘‘Sporozoa’’: Apicomplexa, Microsporidia, Haplosporidia, Paramyxea, Myxosporidia, and Actinosporidia. In Microscopic anatomy of invertebrates, vol. 1, F. W. Harrison and J. O. Corliss (eds.). Wiley-Liss, New York, New York, p. 261–331. ———. 2000. Phylum Haplosporidia Caullery and Mesnil, 1899. In Illustrated guide to Protozoa, 2nd ed., J. J. Lee, G. F. Leedale, and P. Bradbury (eds.). Society of Protozoologists, Lawrence, Kansas, p. 1328–1341. REECE, K. S., M. E. SIDDALL, E. M. BURRESON, AND J. E. GRAVES. 1997. Phylogenetic analysis of Perkinsus based upon actin gene sequences. Journal of Parasitology 83: 417–423. ———, AND N. A. STOKES. 2003. Molecular analysis of a haplosporidian parasite from cultured New Zealand abalone Haliotis iris. Diseases of Aquatic Organisms 53: 61–66. RENAULT, T., N. A. STOKES, B. CHOLLET, N. COCHENNEC, F. BERTHE, AND E. M. BURRESON. 2000. Haplosporidiosis in the Pacific oyster, Crassostrea gigas, from the French Atlantic coast. Diseases of Aquatic Organisms 42: 207–214. SIDDALL, M. E., N. A. STOKES, AND E. M. BURRESON. 1995. Molecular phylogenetic evidence that the phylum Haplosporidia has an alveolate ancestry. Molecular Biology and Evolution 12: 573–581. SPRAGUE, V. 1979. Classification of the Haplosporidia. Marine Fisheries Review 41: 40–44. SWOFFORD, D. L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 10. Sinauer Associates, Sunderland, Massachusetts. THOMPSON, J. D., D. G. HIGGINS, AND T. J. GIBSON. 1994. Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Research 22: 4673–4680.